***generating variables*
*information on running operations, administration and politicians
gen v029driftadm=v02909

gen v029driftpol2=0
replace v029driftpol2=1 if v02907==1
replace v029driftpol2=1 if v02908==1

*information on goals, administration and politicians
gen v029maaladm=v02914
gen v029maalpol2=0
replace v029maalpol2=1 if v02912==1
replace v029maalpol2=1 if v02913==1

*discussion of goals, administration and politicians 
gen v006adm=v0064
gen v006pol2=(v0061+v0062)/2

*making ln-variables of number of SOEs, number of employees, income, fragmentation
gen lnTotal=ln(Total)
gen lnansatt=ln(ansatt2014)
gen lnnetdrcap15=ln(netdrcap15)
gen lnhhi2011=ln(hhi2011)

*ANALYSIS

*distribution and skewness
drop if nonsurvey==0

tab v029maalpol2 v029driftpol2
summarize v029maalpol2 , detail

tab v029maaladm
summarize v029maaladm, detail

tab v029driftpol2
summarize v029driftpol2 , detail

tab v029driftadm
summarize v029driftadm, detail

tab v006pol2
summarize v006pol2, detail

tab v006adm
summarize v006adm, detail

tab v04701
summarize v04701, detail

tab lnansatt
summarize lnansatt, detail

tab Total_kat
summarize Total_kat, detail

tab netdrcap15
summarize netdrcap15, detail

tab hhi2011
summarize hhi2011, detail

tab lnnetdrcap15
summarize lnnetdrcap15, detail

tab lnTotal
summarize lnTotal, detail

*MODEL

ologit v03802 i.v029maalpol2 i.v029maaladm i.v029driftpol2 i.v029driftadm v006pol2 v006adm v04701 lnansatt lnTotal netdrcap15 lnhhi2011, robust or

*checking for multicolinearity (estat vif)
regress v03802 i.v029maalpol2 i.v029maaladm i.v029driftpol2 i.v029driftadm v006pol2 v006adm v04701 lnansatt lnTotal netdrcap15 lnhhi2011, robust
estat vif

*ONLINE APPENDIX
*testing for reversed causality

*generating variable for tenure of CO
gen RådmannAnsattLenge=2014-RådmannAnsattÅr

*generating dummy for new CO and CO with longer tenure
gen RådmannAnsattLenge3=0
replace RådmannAnsattLenge3=1 if RådmannAnsattLenge>1
tab RådmannAnsattLenge3

*correlation beetween CO tenure and independent variables
pwcorr RådmannAnsattLenge3 v029maalpol2, obs sig
pwcorr RådmannAnsattLenge3 v029maaladm, obs sig
pwcorr RådmannAnsattLenge3 v029driftpol2, obs sig
pwcorr RådmannAnsattLenge3 v029driftadm, obs sig
pwcorr RådmannAnsattLenge3 v006pol2, obs sig
pwcorr RådmannAnsattLenge3 v006adm, obs sig
pwcorr RådmannAnsattLenge3 v04701, obs sig


*model run on sample with only new COs and with COs with longer tenure
bysort RådmannAnsattLenge3: ologit v03802 i.v029maalpol2 i.v029maaladm i.v029driftpol2 i.v029driftadm v006pol2 v006adm v04701 lnansatt lnTotal netdrcap15 lnhhi2011, robust or

*model, with tenure as control
ologit v03802 RådmannAnsattLenge3 i.v029maalpol2 i.v029maaladm i.v029driftpol2 i.v029driftadm v006pol2 v006adm v04701 lnansatt lnTotal netdrcap15 lnhhi2011, robust or
ologit v03802 RådmannAnsattLenge i.v029maalpol2 i.v029maaladm i.v029driftpol2 i.v029driftadm v006pol2 v006adm v04701 lnansatt lnTotal netdrcap15 lnhhi2011, robust or
